Global interannual variability in terrestrial ecosystems: sources and spatial distribution using MODIS-derived vegetation indices, social and biophysical factors

Abstract

Variations in global vegetation activity were measured at a global scale, from 2000 to 2006, based on the Enhanced Vegetation Index (EVI) extracted from the 1km resolution Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data. Interannual variations in phenology and/or in annual integrated vegetation index are mapped using change metrics. The relationships between interannual variability and climate, ecosystem disturbances and land use are also examined. Around 14% of the study area experienced high interannual variability in land surface attributes over the six years. The ecosystems most subjected to large fluctuations in surface conditions were the boreal ecosystems, temperate ecozones, and subtropical and tropical steppes. These changes were largely related to rainfall variability and were associated with mean annual rainfall, agriculture, fire regimes and population density. Large population concentrations were mostly found in more stable ecozones. Rainfall and natural fire regimes explained more than half of the land surface variability in Australia. An additional global analysis on trends in vegetation activity also shows that around 4.5% of the vegetated surface of the Earth, excluding deserts and frequently cloudy regions, experienced a continuous decrease in vegetation activity over the six years. This represents more than twice the area experiencing a greening trend over this time period and concerns mostly tropical, subtropical and temperate forest ecozones. The total change in vegetation activity at a global scale and per year amounted to –2% of the annual integrated EVI aggregated across all ecosystems of the study area, on average for the years 2001–2006.

Acknowledgements

This work was funded by a research program on satellite remote sensing for the Belgian Science Policy Office (GLOVEG) and the FP5’s Environment and Sustainable Development Programme of the European Commission (CYCLOPES). Pedram Rowhani gratefully acknowledges the financial support of the Ministry of Research of the Grand-Duchy of Luxembourg (‘Bourse Formation Recherche’). We are grateful to Calli Jenkerson and colleagues from the Land Processes Distributed Active Archive Center for rapid data delivery.